Sulforaphane exhibits antiviral activity against pandemic SARS-CoV-2 and seasonal HCoV-OC43 coronaviruses in vitro and in mice

The COVID-19 pandemic has caused major global morbidity and mortality. Although effective vaccines exist, many individuals remain at risk due to limited access, vaccine hesitancy, or suboptimal immune responses. Available therapies include monoclonal antibodies, convalescent plasma, antivirals, and immunomodulators, with early intervention providing the most benefit. Among oral antivirals, only molnupiravir and ritonavir-boosted nirmatrelvir were authorized at the time. There is a need for safe, easily administered, widely available therapeutics. Sulforaphane (SFN), a phytochemical derived from glucoraphanin in cruciferous vegetables, is a potent activator of NRF2 with antioxidant and anti-inflammatory properties, previously shown to enhance alveolar macrophage function and reduce lung inflammation in preclinical models. The authors hypothesized that SFN could exert antiviral activity against human coronaviruses, including SARS-CoV-2 and HCoV-OC43, and evaluated SFN in vitro and in vivo, including potential synergy with remdesivir.

The paper contextualizes current COVID-19 treatments (monoclonal antibodies, convalescent plasma, antivirals, immunomodulators) and emphasizes the need for additional oral therapeutics. It summarizes SFN’s established biological activities: activation of NRF2, antioxidant and anti-inflammatory effects, enhancement of alveolar macrophage phagocytosis, and prior antiviral activity against influenza and respiratory syncytial virus in preclinical models. The authors note reports that NRF2-dependent genes are suppressed in SARS-CoV-2 infection and that NRF2 agonists (e.g., dimethyl fumarate, 4-oxalyl-lactate) can inhibit SARS-CoV-2 replication in vitro. They also discuss mechanistic links whereby NRF2 activation may reduce ACE2 expression and modulate proinflammatory cytokines (IL-6, IL-8), suggesting a rationale for NRF2-targeted host-directed therapy for COVID-19.

In vitro studies: The antiviral activity of SFN was assessed in Vero C1008 cells (African green monkey kidney), human fibroblasts (MRC-5), and human intestinal epithelial cells (Caco-2). Cells were exposed to SFN 1–2 h prior to inoculation with HCoV-OC43 or SARS-CoV-2 (Wuhan-Hu-1 and additional strains/variants), or 24 h post-inoculation to assess effects on established infection. Dose-response curves were generated; IC50 and therapeutic index (TI) were calculated. Cytotoxicity (TD50) was determined. Additional SARS-CoV-2 strains tested included USA/VA2020, clinical 614G+ isolates (USA/MDH-20/200; USA/DCHP-71220D), and variants Delta (B.1.617.2) and Omicron. Drug synergy with remdesivir was evaluated via two-drug combination assays with combination index (CI) analysis (isobolograms). To probe mechanism, NRF2 knockdown Caco-2 cells were generated via CRISPR/Cas9; infected cells were treated with SFN and viral RNA quantified by RT-qPCR to assess NRF2 dependence. Viral RNA in supernatants and tissues was measured using standard extraction (Zymo Quick-RNA Viral Kit), cDNA synthesis, and RT-qPCR targeting the SARS-CoV-2 N gene with specified cycling parameters and normalization (Rpl02a for tissue). In vivo studies: Male K18-hACE2 transgenic mice were intranasally inoculated with 8×10^5 CID50 SARS-CoV-2 (USA/VA2020). SFN (30 mg/kg/day, oral gavage) was administered prophylactically starting one day before infection and continued daily. Outcomes included body weight change, bronchoalveolar lavage (BAL) total protein as a marker of lung injury, viral burden in BAL and lung (qPCR and titers), lung histopathology (H&E, semi-quantitative scoring), and immunohistochemistry for SARS-CoV-2 spike protein with quantitative image analysis. High-dimensional flow cytometry with UMAP visualization assessed immune cell populations in lung and spleen, including myeloid cell recruitment and T cell activation/cytokine production. Statistical analyses used GraphPad Prism; significance at P<0.05.

- SFN inhibited virus-induced cytopathic effects and replication in vitro: - HCoV-OC43 in Vero C1008 cells: IC50 ≈ 1.0 µM (95% CI 0.71–2.0); TI ≈ 5. - HCoV-OC43 in human MRC-5 cells: IC50 ≈ 18 µM (95% CI 9.7–35); TI ≈ 5; TD50 ≈ 73–89 µM. - SARS-CoV-2 Wuhan-Hu-1 in Vero C1008 cells (pre-treatment): IC50 ≈ 12 µM (95% CI 4.7–30); TI ≈ 7. - SARS-CoV-2 Wuhan-Hu-1 in Vero C1008 cells (added 24 h post-infection): IC50 ≈ 1.3 µM (lower micromolar range reported). - SARS-CoV-2 in Caco-2 cells: IC50 ≈ 4.9 µM (qPCR readout; pre-treatment). - Additional strains/variants: USA/VA2020 IC50 ≈ 28 µM (95% CI 14.7–46.4); clinical 614G isolates IC50 ≈ 29 µM (95% CI 8.2–102.3); Delta IC50 ≈ 5.6 µM (95% CI 4.1–7.8); Omicron IC50 ≈ 3.3 µM (95% CI 0.9–11.8). - Post-infection activity: SFN inhibited established HCoV-OC43 (IC50 ≈ 1.8 µM; 95% CI 1.4–4.1) and SARS-CoV-2 infections when added 24 h after inoculation. - Synergy with remdesivir: Two-drug combination assays showed synergistic inhibition (CI<1) of HCoV-OC43 and SARS-CoV-2 at sub-IC50 concentrations for each drug. - Mechanism: NRF2 knockdown did not abrogate SFN’s in vitro antiviral effect in Caco-2 cells, suggesting NRF2-independent antiviral activity under these conditions. - In vivo efficacy in K18-hACE2 mice (prophylactic SFN, 30 mg/kg/day): - Less weight loss vs untreated infected controls (P<0.001). - Lower BAL total protein, indicating reduced lung injury (P<0.0001). - Lower viral burden in BAL (P=0.04) and approximately 1.5-log reduction in lung viral titers (P=0.004). - Reduced lung pathology: lower histopathology scores (e.g., 6/16 in treated; P=0.0008), with less alveolar and peribronchial inflammation. - Immunostaining: markedly reduced lung area positive for SARS-CoV-2 spike protein; untreated lungs had ~4.4-fold higher virus-associated area vs SFN-treated (P=0.01). - Immune modulation: decreased recruitment of myeloid cells (monocytes, dendritic cells) to lung, reduced T cell activation and cytokine production; systemic (spleen) changes were minimal.

The study demonstrates that sulforaphane exerts both direct antiviral and immunomodulatory effects against human coronaviruses. SFN inhibited replication of multiple SARS-CoV-2 strains and variants (including Delta and Omicron) and HCoV-OC43 in vitro, showed synergy with remdesivir, and retained activity when administered after infection onset. In mice, prophylactic SFN reduced viral loads, lung injury, and pathology, and tempered excessive pulmonary immune activation while preserving antiviral responses. Mechanistically, while SFN is a potent NRF2 activator, NRF2 knockdown did not eliminate its in vitro antiviral effect, indicating NRF2-independent antiviral actions alongside NRF2-mediated anti-inflammatory effects. Clinically, reported human SFN exposures (Cmax ~1.9 µM after SFN-rich sprouts) overlap with in vitro IC50 values in human cells, supporting translational potential. By targeting host pathways that modulate both viral replication and inflammatory injury, SFN may complement existing antivirals and help prevent progression to severe COVID-19.

SFN inhibits replication of SARS-CoV-2 and HCoV-OC43 in vitro and reduces viral burden, lung injury, and inflammatory pathology in a mouse model of SARS-CoV-2 infection. It also synergizes with remdesivir and modulates detrimental immune responses. Given its oral bioavailability, known safety profile, and pharmacologically relevant exposures, SFN is a promising candidate for prevention and treatment of COVID-19 and potentially other respiratory viral infections. Future work should include controlled clinical trials to define efficacy, optimal dosing and formulation, pharmacokinetics/pharmacodynamics in diverse populations, timing relative to infection, and mechanistic studies to delineate NRF2-dependent and independent pathways and combination strategies with approved antivirals.

- Mouse model limitations: K18-hACE2 transgene expression is non-physiological with tissue expression patterns differing from endogenous ACE2, which may affect disease manifestations and drug responses. - Sex bias: Only male mice were used, precluding assessment of sex-specific effects. - Variable SFN exposure: Oral SFN absorption and bioavailability are influenced by intestinal microflora and preparation, potentially introducing inter-animal variability. - In vitro model constraints: Some assays used non-human cell lines (Vero cells) and may not fully reflect human respiratory epithelium; cytopathic readouts differ by cell type (e.g., limited cytopathogenicity in human cells). - Mechanistic ambiguity: Although NRF2 knockdown did not abolish antiviral effects, the relative contributions of NRF2-dependent vs NRF2-independent pathways remain unresolved.